|
PopZeus posted:feeling a little crazy - was there a video that was very close-up of Elizabeth Holmes where she is creepily talking right into camera? google is failing me, or maybe i'm thinking of another tech person making a similar video? I love how apparently at one point she was advised "Hey, you should try to find a unique aesthetic style for your public appearances. Steve Jobs did and it worked well for him", and then she proceeded to just outright copy Steve Jobs' turtleneck deal. 
|
#
¿
Feb 19, 2021 23:09
|
|
|
|
| # ¿ May 8, 2024 22:20 |
|
|
Volmarias posted:Holy poo poo, I knew it was high, I didn't realize it was that high! IIRC at this point Bitcoin uses up about as much electricity as literally the entire conventional global banking system. More than entire actual countries.
|
|
#
¿
Apr 13, 2021 21:16
|
|
|
What the gently caress. That's like, on the level of a decently impressive tech demo of something in ongoing development. Why in the world is that allowed on actual roads with actual people in the wild?!  It's an old truism that the intermediate levels autonomous driving along the lines of "can drive on its own but driver needs to be able to react within X seconds" are essentially worthless, since to be able to do that safely you basically need the same capabilities as 100% safe reliable autonomous driving. Tesla just manages to prove that again by just doing it unsafely.
|
|
#
¿
Apr 28, 2021 08:46
|
|
|
CommieGIR posted:A good anecdote: Not quite mechanical engineering, but I was hired by a development firm to teach Programmers, often with degrees in Computer Science, how to properly manage Linux systems and develop for them. They just didn't know very well, it wasn't covered in their computer classes. They were incredibly skilled at the languages they mastered in school, they did know the basics, but they didn't know how to properly use *nix systems. Yeah, that's pretty much how it went for me. Came out of uni with a pretty solid grasp of Java, C++, Matlab, even loving Assembler, but barely a lick of experience using linux. When I started my first job, it involved a lot of circumspect googling of even the most rudimentary stuff like "how do I ssh to places". Though in fairness, "how to google poo poo and apply it quickly" is a pretty fundamental skillset for developers anyway. 
|
|
#
¿
Aug 29, 2021 20:01
|
|
|
Blut posted:And obviously videos like the one in my post speak for themselves. Those protestors don't look reasonable with their actions/behaviour. Or with physically storming into senior staff meetings in Netflix, leaking confidential internal financial data etc. Surely theres a more adult way to have a conversation about workplace issues? This particular bit didn't happen. It was a digital meeting and the employee in question was provided the link by a director, which is obviously an implicit invitation to attend. She's since been reinstated: https://twitter.com/rainofterra/status/1448097310026747908
|
|
#
¿
Oct 22, 2021 12:12
|
|
|
On the topic of the bad tunnel: https://twitter.com/adamtranter/status/1479149037253238790
|
|
#
¿
Jan 6, 2022 20:57
|
|
|
Blut posted:How can the AI not manage to drive in one of those? It looks like the ideal environment for an AI - no pedestrians, intersections, animals, weather... Basically just stop and go and follow the gentle curves. The funny thing is that other organizations actually do manage that exact thing already, and better than Tesla. In my city we just had a pilot project with an autonomous bus servicing a short circular route as a practice run. It was a simple route in a fairly calm part of the city, but it was on otherwise uncontrolled public streets with cars and pedestrians both. It turned out to be a solid success, with the autonomous bus managing to navigate daily traffic without any incidents for a year. So even if you're dead set on having your public transport in tiny autonomous vehicles along predetermined routes, you still don't actually need to bother with the drat tunnels.
|
|
#
¿
Jan 6, 2022 23:26
|
|
|
withak posted:One of our marketing people likes to say that most of the features of Salesforce are targeted at helping you justify the next renewal of your Salesforce subscription. Yeah, that's a common theme with combined systems like this, it's similar with stuff like SAP. You've got a fairly aggressive marketing offering lots of attractive up-front offers and support, all geared to get you to put all your data into it. Once they got you to that point you're pretty much locked in, because actually extracting all your data and putting it into some other system would be even more cost and effort than just keeping your subscription going. Also yeah, Teams is a giant pain in the rear end. It took us weeks to even get it to work because Microsoft kept trying to log people in with the wrong accounts, and we still can't use the calendar function across organizations. Just a complete failure all around.
|
|
#
¿
Mar 24, 2022 13:45
|
|
|
Yeah, there are a lot of cheeky attempts to try and trick people into giving consent against their will. Especially early on right after the passage of the GDPR, some cheeky fuckers only gave you a list of their 100+ advertisers and you'd have to manually click and opt out of them one at a time. However, the legal guidelines are actually quite clear on it: opting out must be as quick and easy to do as opting in, and obviously cannot be deceptive. Most of the common plugins manage it at this point, but there are still some that expect you to go through several screens to opt out. And of course there are still some US sites that outright go "Well if you don't let us dump a hundred trackers on you, we just won't let you use our site at all!"
|
|
#
¿
Apr 22, 2022 08:04
|
|
|
BiggerBoat posted:I've gotten one or two calls from my own number so that's really something. The European General Data Protection Regulation: https://en.wikipedia.org/wiki/General_Data_Protection_Regulation In short, it's supposed to significantly constrain what data companies are allowed to collect about users and to what degree they may process it, and backs it up with actually quite beefy fines. For example, companies are no longer allowed to just send you marketing emails out of the cold, you need to either have a pre-existing business relation or have explicitly consented to receiving those emails. Similarly, every such email must have a link to unsubscribe you from whichever list you're on, and that function must actually work. While there's certainly quite a bit to criticize about the gdpr's actual execution, it sure does have a point.
|
|
#
¿
Jul 14, 2022 09:09
|
|
|
Blut posted:Investing in a $200 "dumb" laser printer is something I recommend to everyone. Its just so much easier in the long run than dealing with crappy HP or whatever inkjets. Yeah, I got some old Laserjet that's like 15 years old at this point and still runs like a charm. drat thing is indestructible, easy to use, and the toner lasts forever.
|
|
#
¿
Jul 22, 2022 15:29
|
|
|
AvesPKS posted:Why does a chess playing robot need the strength of 3 people Yeah, it's been a long while since I played around with one of those, but I'm pretty sure they generally have sensors telling you how much force/torque they're currently exerting, and consequently let you limit the maximum force you want to use.
|
|
#
¿
Jul 24, 2022 15:00
|
|
|
Yeah, there was a whole thing a while back where Apple built in a function that lets you easily opt out of all sorts of trackers (except their own, of course), and it ended up costing Meta literally ten billion in the span of just one year. That, combined with similar tracking prevention measures like the GDPR and certain browsers, really murdered the hell out of Meta's business model. No wonder they're trying so hard to promote their whole metaverse thing.
|
|
#
¿
Jul 31, 2022 12:10
|
|
|
As somebody who does own a VR set, I figure ergonomics are going to be the main barrier. As much as I enjoy playing VR games, it's always kind of a production to dig out that big old headset, you get sweaty underneath, and after an hour or thereabouts you'll start feeling the pressure on your scalp. Enjoyable VR experience or not, having half a kilo or more strapped to the front of your face is just plain physically unpleasant in the longer term. If and when they manage to shrink a headset down to about the form factor of about a pair of snow or diving goggles, that's where I'd expect the actual mainstream breakthrough to come.
|
|
#
¿
Aug 19, 2022 16:49
|
|
|
Yeah, it very much reeks of startup/entrepreneur brain where you assume that literally everyone else is just too stuck in their ways and afraid to innovate, only to run headfirst into hundreds of problems that literally everyone else has already encountered and solved before. It's doubly funny that it seems to happen in a software context. Basically the first thing I learned in practical development was: "No matter what you're trying to do, somebody else likely already tried it, and probably did it better than you would. Do your loving research and don't waste your time on solved problems".
|
|
#
¿
Sep 23, 2022 20:28
|
|
|
Jaxyon posted:How's twitter going It's an incredible encapsulation of upper management brain. Create a big old waste of everyone's time and effort, just so that they can convince themselves that they have any purpose whatsoever.
|
|
#
¿
Oct 29, 2022 23:49
|
|
|
Karia posted:It baffles me that the people responding to that tweet are objecting to this because it's an inefficient way to review code and a terrible metric for judging productivity. Sure, that's true, but the really weird thing is that Musk feels the need to review the code-base at all. Twitter's code is fine! It's like someone buying a restaurant that's known to have a lovely menu, and then day 1 he comes in and says "I want to look at how you wash your dishes." Sure, dishwashing is important and necessary for the restaurant to work, but they've got dishwashers who know what they're doing already, and the new owner clearly has more important things to look at. Policy is the problem, and coders aren't setting policy! If he wants a new feature, he just has to tell someone and it'll happen! I'm betting it's purely for the benefit of Musk's own ego. He still likes to pretend he's a real-life Tony Stark who actually has a deep understanding of the actual technologies involved. So now he'll have a parade of developers in and out of his office, each trying their damnedest to give this idiot the vaguest understanding of unrelated code snippets in the span of ten minuts. All so that when he gets a high-level presentation about feature X in the future, he can nod sagely and go "Ah yes, I've seen the code for that", while everybody rolls their eyes and sighs in exasperation.
|
|
#
¿
Oct 30, 2022 09:35
|
|
|
W-why would you want to look at vine code in the first place? Twitter already has a wholeass infrastructure for uploading, hosting, and playing video files, so if you did want to make a pseudo-vine or tiktok you'd likely be better off building on that instead. The thing that Musk doesn't seem to understand is that the actual writing of the code is usually the smallest part of adding a new feature. The trickier and generally more important parts are properly defining what exactly you want/need to do (and whether that's of any use), understanding how it would interact with other parts of your system, and how to present it to your users. Plus of course things that aren't directly related to the functionality like visual design, marketing, translations, and possibly legal matters. Once all that is settled, getting a dev team or three to sit down and then write the code that makes things happen is a relatively straightforward matter. Perestroika fucked around with this message at 22:39 on Oct 31, 2022 |
|
#
¿
Oct 31, 2022 22:19
|
|
|
Anno posted:Isn’t that a replica of the .357 revolver from the new Deus Ex games? Feels like there should probably be some commentary there somewhere. It's seems to be specifically a replica of a Team Fortress 2 gun patterned after that Deux Ex gun: https://wiki.teamfortress.com/wiki/Diamondback. And not a particularly well-made one, either.
|
|
#
¿
Nov 28, 2022 18:49
|
|
|
dr_rat posted:Also that, oh "we implemented that feature in five minutes". It's kinda funny, back when I was in uni and Agile development was the new hotness, this was the exact thing we were warned against. The general idea with agile is that you quickly slap together a barely working prototype to see if you're on the right track, and then iterate on that based on feedback. And I still recall almost word for word the "downside" section of that approach: "The greatest risk with agile development is managers believing it can be used to achieve finished products more quickly, when that is explicitly not its purpose". Yes, you can get something that technically works quite quickly, but you still need to do the actual legwork of validating it, technical cleanup, documentation, and so on. Trying to skip those steps is a typical pitfall of startups, but to see an actual multi-billion company going that way is just .
|
|
#
¿
Dec 3, 2022 12:03
|
|
|
Yeah, we used to go whole hog on Jira for a while as well, subdividing every new development into cards with estimations and everything. But eventually it turned out that just putting our heads together over a shared google doc and jotting down an outline and some relevant notes worked just as well and took like a tenth of the time. Then again we are a pretty small team, so it might have more utility with more people involved.
|
|
#
¿
Jan 1, 2023 12:40
|
|
|
It would literally be more efficient (and safer) if they'd just let people walk through the tunnel. It's not even two miles long.
|
|
#
¿
Jan 8, 2023 19:25
|
|
|
Boris Galerkin posted:I'm not sure how these obviously cherry-picked examples of ChatGPT being hilariously wrong are the owns that people seem to think they are. I'm 100% positive that for every cherry-picked bad example there's a "holy poo poo this is actually genius" good example out there. They illustrate the core issue/misunderstanding people have with ChatGPT: It doesn't have any understanding of the things it outputs, it is purely an association machine. Yes, if you feed it with enough data and sufficiently refine it, it can be correct an impressive amount of the time by sheer weight of numbers. But there's not even the most low-level sanity check that it's not completely wrong about even the most basic of facts it puts out, and the way it dresses up the answers in fairly verbose texts tends to obscure that facet. You can already see the first idiots going "Well I asked the Most Advanced AI Of All Time about the thing we're arguing, and it totally said I was correct, so there!". Yes, it can point you in the right direction a lot of the time. But as it stands cannot be considered reliable to the same degree as a manually curated knowledge base.
|
|
#
¿
Jan 28, 2023 19:06
|
|
|
One factor is also just the sheer scale of the data being fed into this one. IIRC the first iteration of ChatGPT was trained on "only" some 5 GB of text data, meanwhile the most recent one is closer to about 600 GB. Similarly, the number of parameters went up from a few million to hundreds of billions and may soon go into the trillions. So to a degree it's also a matter of throwing vast amounts of resources at it. That gives it enough of a breadth of data to draw from that it'll be able to give a convincing (if not correct) response to a much wider range of prompts than before.
|
|
#
¿
Mar 24, 2023 15:03
|
|
|
Amphigory posted:Has no one realised what page we're on Unfortunately, digital illiteracy has become very common these days. So many kids never learn how to 1337.
|
|
#
¿
Apr 20, 2023 21:03
|
|
|
At first I was a bit confused why removing the operator would be advantageous points-wise, since no signal would/should still result in no strike. But then I realised they probably associated a no-go order with a points penalty, as that would mean the drone incorrectly targeted something. But that in turns poses the question why in the world the operator would give the go-order on friendly targets, or indeed themself. Though of course it's just as possible that the thing started targeting friendly infrastructure for unrelated reasons (e.g. "bomb everything" gives more points faster than "bomb things selectively") and the guy projected a narrative on top of that, people are good at that. e: Also on that topic, this is a pretty fun list of similar specification errors: https://docs.google.com/spreadsheets/u/1/d/e/2PACX-1vRPiprOaC3HsCf5Tuum8bRfzYUiKLRqJmbOoC-32JorNdfyTiRRsR7Ea5eWtvsWzuxo8bjOxCG84dAg/pubhtml
|
|
#
¿
Jun 2, 2023 08:44
|
|
|
Also, a lot of AI stuff still fundamentally boils down to vectors in one fashion or another, and GPUs are generally very good at handling vector math.
|
|
#
¿
Jun 6, 2023 09:55
|
|
|
Fundamentally the Apple VR shares the same problem as all others do: There's not really anything to use them for yet. Meta is still trying to go all-in on some sort of virtual workspace thing, and is learning the hard way that nobody wants that. Others go mostly for games, which works but is still fairly niche, and something where Apple falls down out of the gate what with not even having controllers. Just about everything they've shown boils down to "things you can already do, but now with an uncomfortable piece of poo poo strapped to your face and a worse interface". Office work you can do better with a traditional screen or three, online meetings you can do better with a webcam. It has neither the convenience of something like a smart phone or watch, nor does it yet offer much of anything unique you'd genuinely need or want it for. If they want it for AR stuff and spontaneous use like a wearable, it'd need to be drastically more lightweight with a simpler form-factor. For a full-immersion VR system, it just plain lacks a core use-case. Perestroika fucked around with this message at 15:35 on Jun 6, 2023 |
|
#
¿
Jun 6, 2023 15:17
|
|
|
Boris Galerkin posted:Pay me 300k base + 300k stock and I'll gladly rent and live in a shithole for 3 years before moving to a non-shithole and using the experience as a jumping off point to demand a remote only job. Hell, even if you end up paying rents to the tune of 5k/month (which google tells me would be a bit above the median for a 2 bedroom), that should still leave you with some 120k after taxes. Even with high costs of living otherwise, I don't see how this wouldn't leave you with more disposable income than most people have in gross income.
|
|
#
¿
Jun 25, 2023 08:40
|
|
|
Another somewhat common arrangement I've been seeing in Germany is having a washing machine in your own apartment (usually kitchen, sometimes bathroom, rarely a separate little closet), but also having a shared, well-ventilated drying room in the cellar or attic where you can hang your stuff up to dry. The idea being that air-drying in your own apartment could lead to mold growth, especially in winter when people might be less likely to ensure proper ventilation.
|
|
#
¿
Jul 17, 2023 12:21
|
|
|
Yeah, I'm leaning towards this mostly being a marketing ploy as well. "Why did your organization have a huge leadership kerfluffle recently? Uuuuh because our product is just too good, you see".
|
|
#
¿
Nov 25, 2023 09:44
|
|
|
That's impressively terrible. At my last job we've built a web app that also involves user management and authentication, and despite having only all of 3 developers with not too much experience in that regard we apparently managed something a whole lot more secure than that, including email opt-in as well as revokable session tokens. And not because we're particularly clever or anything, but rather because that's extremely well-trodden ground with a whole host of established best practices and available packages. And yet here is OpenAI with practically infinite budget and they're loving up something this basic
|
|
#
¿
Dec 11, 2023 15:35
|
|
|
Clarste posted:Employers thinking it's B when it's actually A lets us have both problems at once! Which is why the hyping of AI is the most dangerous problem, regardless of anything else. Exactly. I think it's pretty uncontroversial to say that the current state of generative AI isn't at a point where it matches or even surpasses the work of competent human artists. And it will be a good while yet to get there, if it ever does. But it's extremely possible that dumbass execs will be sold on the idea that it actually already is, and we could end up with the worst of both worlds where a lot of great artists end up out of work while much of readily accessible "art" becomes a lot worse.
|
|
#
¿
Jan 22, 2024 16:30
|
|
|
Leon Trotsky 2012 posted:If people have the choice of a book with lovely AI cover art or a book with an image from a human artist, but it doesn't make a difference in what book they prefer, then did the "good" cover art really matter in the first place? I'd absolutely say yes. For the reader, it matters because now their book looks a little poo poo on their shelf. It's a little less fun to look at, they may be a little less likely to pick it back up even if they like the actual content. It's a fairly minor difference, but nonzero. For the author, it's a bit bigger. For one there's the emotional aspect of having created something you're really proud of, only to have a shitass cover slapped on it by your publisher. More practically, it may also limit the overall reach of their book, as good cover art can be a great tool to create interest and communicate the overall theme and vibe of a book. To the artist who would have made that cover, it's the biggest loss. Before, they might have been able to finance their art full-time through jobs like this, and now might be forced to pursue it part-time next to a regular job, creating less art overall (or even none at all). That's a significant blow to them, and honestly to humanity as a whole. Individually, these may seem very minor. But multiplied over however many books, readers, authors, and artists, it adds up. More fundamentally, it just works out to everything getting a little worse for everyone involved, with the only beneficiaries being the publishers getting to show slightly higher profits. Perestroika fucked around with this message at 18:52 on Jan 22, 2024 |
|
#
¿
Jan 22, 2024 17:56
|
|
|
Somewhat related to the amazon enshittification, it's insanely annoying how reliable product review sites are borderline dead now, and it's all their fault. I'm looking for a somewhat specific item and I'm trying to get an idea of the overall range of quality available. I guarantee you, some 10 years ago there would have been an nice niche blog or site around where somebody would have written up a nice explanation of the relevant factors, and then a comparison of whatever's currently on the market, written by somebody who had probably handled some or all of those things personally. These days, though? Just about every comparison site is just some meaningless (probably partially generated) blurb followed by a lovely table that's just the Amazon search results formatted a little differently (with affiliate links, naturally).
|
|
#
¿
Apr 3, 2024 08:35
|
|
|
Yeah I've switched to DDG a while back as well, and it's text* search is generally fair bit better. But even then some of the damage that google has done remains, as once-useful resources have since been shuttered or converted to SEO slurry because they couldn't keep up. *Its image search is kinda funny cause it seems like the moment your search has even one risque word in it it's just rows and rows of porn results.
|
|
#
¿
Apr 3, 2024 12:27
|
|
|
|
| # ¿ May 8, 2024 22:20 |
|
|
PhazonLink posted:was just about to ask about just normal rain. Supposedly it just puts the car into neutral, disables the wipers, and locks the charging port closed. It does absolutely nothing to actually prevent water from going anywhere.
|
|
#
¿
Apr 26, 2024 08:34
|
|